AMCS, A Glazov, V Radescu, S Whitehead, A Sapronov Confronting HERAPDF with Tevatron data and LHC data PDF4LHC march 7th 2011 A M Cooper-Sarkar +Tevatron W-asymmetry and Z0 rapidity spectrum +Tevatron jets--- with apologies to D0, so far only CDF data fitted LHCb lepton asymmetry CMS lepton asymmetry ATLAS lepton asymmetry AMCS, A Glazov, V Radescu, S Whitehead, A Sapronov

How good or bad is the HERAPDF description of the Tevatron data?. We usually just show you plots. Let’s calculate χ2- and then FIT the data. Can the error bands of HERAPDF!.5 cover the variation due to input of new data? The description of the CDF Z0 rapidity LOOKS OK but what is the χ2 for these data points to HERAPDF1.5 central values? 36/28 -acceptable But this does NOT account for the error band of the HERAPDF fit

So FIT the CDF Z0 rapidity data: χ2/ndp =30.5/28 And look to see if the resulting PDF has shifted outside error bands. Here the HERADF+CDF Z0 rapidity fit is the blue line– it does NOT move outside the HERAPDF1.5 error bands The fit is done with HERAPDF standard parametrisation BUT with the low-x Bdv parameter of the d-valence freed so that Bdv ≠ Buv (note this was one of our standard parametrisation variations)

The description of the CDF W-asymmetry LOOKS OK but what is the χ2 for these data points to HERAPDF 1.5 central values? 40/13 --not very good But this does NOT account for the error band of the fit

So FIT the CDF W-asymmetry data: χ2 = 21/13 And look to see if the resulting PDF has shifted outside our error bands. Here the HERADF+CDF W- asymmetry fit is the blue line– it does NOT move outside the HERAPDF1.5 error bands Note the χ2 is comparable to that of NNPDF 24/13 The fit is done with HERAPDF standard parametrisation BUT with the low-x Bdv parameter of the d-valence freed so that Bdv ≠ Buv Since the W-asymmetry measures uv-dv is is desirable to allow more freedom in the valence sector (note this was one of our standard parametrisation variations)

Sticking with W and Z data move to the LHC HERAPDF1.0 and 1.5 compared to LHCb lepton asymmetry data- no time to fit yet From G. Watt

HERAPDF1.5 compared to CMS lepton asymmetry data- go the data points from La Thuile no time to fit yet- but it looks pretty good Pt lepton > 25 GeV? Also Etmiss > 25 GeV?

HERAPDF1.5 compared to ATLAS muon asymmetry data– this I have had time to fit(!) What is the χ2 for these data points to HERAPDF 1.5 central values? 27/11 --not very good But this does NOT account for the error band of the fit Ptmuon > 20 GeV Et miss > 25 GeV

So FIT the ATLAS muon asymmetry data: χ2 = 13.5/11 And look to see if the resulting PDF has shifted outside our error bands. Here the HERADF+CDF W- asymmetry fit is the blue line– it does NOT move outside the HERAPDF1.5 error bands The fit is done with HERAPDF standard parametrisation BUT with the low-x Bdv parameter of the d-valence freed so that Bdv ≠ Buv Since the W-asymmetry measures uv-dv is is desirable to allow more freedom in the valence sector (note this was one of our standard parametrisation variations)

Now return to Tevatron jets The description of the D0 jet data LOOKS OK but what is the χ2 for such jet data to HERAPDF1.5 central values?

Table from G Watt The χ2 to these CDF data for HERAPDF1.5 is 176/76 –close to Watt’s value for HERAPDF1.0 – not surprising since high-x gluon does not change much 1.0 to 1.5

After fitting these CDF jet data the χ2= 113/76 And look to see if the resulting PDF has shifted outside our error bands. Here the HERADF+CDF D0 rapidity fit is the blue line– it does NOT move much outside the HERAPDF1.5 error bands- does have a hard high-x gluon- at the edge of the error band. Better χ2 would come from the use of the flexible parametrisation and free alphas… future work

But of course we are working on improved HERAPDF fits The message is that the HERAPDF1.5 describes the Tevatron and early LHC data well – within its error bands But of course we are working on improved HERAPDF fits And NNLO HERAPDF1.5f with uncertainty bands is on its way..

extras

Fit to CDF jets data with very flexible paraemtrisation Chisq= 78/76.. But outside HERAPDf1.5 ! Could also try fitting alphas on less flexible param